Campus stresses 'smart agriculture' in its research

What will farms of the future look like?

The University of California, Davis, created a strategic planning initiative and a competition last year, inviting faculty, staff and students to answer what issues the campus should be addressing during the next decade, and "smart farm" was one of about a dozen ideas selected, according to Daniel Slaughter, a professor in the campus Department of Biological and Agricultural Engineering.

Long term, he said, the idea is to use research to ensure the agricultural production system has the ability to supply healthy, nutritious and flavorful food in the future.

In the short term, Slaughter said there are scientific challenges for agriculture that need to be solved in order to achieve the long-term objectives, including water, fertility, labor and how climate change could affect plant and animal care.

"We also are looking at the transformation that needs to occur in the educational arena, and as the farms of the future change and transform, our training systems and the way that we look at career opportunities for farm laborers need to change also," Slaughter said.

In one part of the university's research program into smart agriculture, Professor Stavros Vougioukas is working on a robot for use in berry fields called FRAIL-bot—short for Fragile Crop Harvest-Aiding Mobile Robots.

Rather than harvesting berries, the FRAIL-bot transports harvested strawberries from the worker to an unloading station. The FRAIL-bots act, in a sense, as a courier service that goes to the worker, picks up the strawberries, and takes them to the unloading station where someone unloads the tray and puts on an empty tray.

The robot will know which workers need trays picked up or brought to them from a sensor the workers will wear. The sensors are also used to monitor their picking rate, so the robots will know when to go to a specific worker.

Workers spend 25-30% of their time transporting the fruit to the unloading station, so using the FRAIL-bots to transport the fruit will increase efficiency.

Besides improving efficiency, the research is also looking at worker health and safety. The robots could tell a worker when they've been in a crouched position too long or when they need to take a break.

Having robots actually harvest the strawberries remains challenging, Slaughter said, mainly due to the robots needing to distinguish berries from foliage, and ripe from unripe berries.

"Also, they're very delicate, so harvesting them delicately is a challenge," he said.

Other research is looking at mechanized orchard-harvesting systems.

For decades, attempts have been made to fully automate harvest of fresh market fruits, but with little success mainly due to the need for a robot to selectively pick the ripe fruit. But a human-robot collaboration—where workers are on a platform and the platform positions them to pick the fruit—could eliminate the need for climbing ladders and therefore improve efficiency.

Cameras would be positioned to determine position of the fruit in the tree, then the platform would automatically position the workers so that they can optimally reach the fruit, Slaughter said.

Several autonomous weeding systems are being researched, and four prototypes have been built.

A current project is called crop signaling, in which plants are treated with a unique, machine-readable compound at planting, then the crop signal is used to differentiate weeds from plants.

Another approach uses a tractor with GPS that records the latitude and longitude of where each seed is planted. Ultimately, a map provides the coordinates of every plant, then the robot reads the map. Using the latitudes and longitudes to show where the plants are, it weeds around them with miniature knives attached to robotic arms.

"So it's hoeing along, and then as it approaches the place where you put a seed or seedling, the knives will move out of the way and around that spot, and come back after you pass that (plant) and continue weeding," Slaughter said.

With the robotic weeding, the actual mechanical part of the project is well developed, he said, but the sensor side on the artificial intelligence needs improvement.

Completely autonomous vehicles are being used in vineyards to apply fungicides, Slaughter said, using a vehicle about the size of an all-terrain vehicle or golf cart that drives up and down the rows in the vineyards.

"It has a smart sprayer on board that accesses the GPS map, and when it gets to each vine it can apply the chemicals that are needed by each vine," he said.

The vehicle turns off the sprayer at the end of the row, makes the turn, then goes down the next row, Slaughter said.

Companies are working on autonomous ATVs that Slaughter said he expects will become available within about a five-year window.

With technologies that have improved labor efficiencies, early research shows initial labor savings of 50% or more, he said.

"The pattern is very consistent with actually the way the first tomato harvesters were introduced in the mid-'60s," Slaughter said.

Initial labor savings using the tomato harvester was about half, he said, and during the next 15 years there were incremental improvements until it reached a 90-95% labor savings.

"We're seeing similar patterns in some of these new technologies," Slaughter said.

"We expect that with continued development improvements in artificial intelligence and sensors to see it to the point where we get closer to that 90% number," he said.

(Kathy Coatney is a reporter in Bend, Ore. She may be contacted at kacoatney@gmail.com.)